A

DamdEEIf4fvv©

K

1

stpcpfdioOtoih

0d

Journal of the European Ceramic Society 27 (2007) 2209–2215

Characterization of rheological properties of colloidal zirconia

B. Jachimska, Z. Adamczyk ∗Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, 30-239 Cracow, Niezapominajek 8, Poland

Received 25 March 2006; received in revised form 12 July 2006; accepted 21 July 2006Available online 22 September 2006

bstract

ynamic viscosity of aqueous suspensions of nanosized zirconia (ZrO2) have been studied for the low volume fraction range. The specific surfacerea of dry powder was determined from the BET method. The zeta potential of zirconia particles as a function of pH was measured by theicroelectrophoretic method. The isoelectric point found in this way was 4.7. The particle density in aqueous suspensions was found by the

ilution method. The dynamic viscosity of suspensions was measured by using a capillary viscometer that eliminated the sedimentation effects.xperimental data showed that for dilute zirconia suspension, the relative viscosity increased more rapidly with the volume fraction than that theinstein formula predicts. This allowed one to calculate the specific hydrodynamic volume of particles in the suspensions and their apparent density.

t was found that particles forming zirconia suspensions were composed of aggregates having porosity of 40–50%. The size of the primary particlesorming these aggregates was 0.2 �m that agrees well with the BET specific surface data. The influence of an anionic polyelectrolyte:polysodium-styrenesulfonate (PSS) on zirconia suspension viscosity also was studied. First the PSS viscosity alone was measured as a function of its volume

raction for various ionic strength of the solutions. The data were interpreted in terms of the flexible rod model of the polyelectrolyte. Then, theiscosity of ZrO2 in PSS solutions of fixed concentration was measured as a function of the concentration of zirconia. It was revealed that theiscosity of the mixtures was proportional to the product of the zirconia and polyelectrolyte viscosities taken separately.

2006 Elsevier Ltd. All rights reserved.

iusrmipa

eocft

eyword: ZrO2; Dispersants; Rheology; Suspensions

. Introduction

Understanding the rheological behavior of colloid suspen-ions is important in the analysis and control of the proper-ies of ceramic materials. High quality and reliability of theseroducts can be achieved by using homogeneous suspensionsomposed of well-dispersed particles. Besides stability of sus-ensions and lack of aggregation processes, a high fluidity is vitalor producing ceramics of complex geometrical shape with aefect-free microstructure after sintering.1–3 In many technolog-cal processes this is achieved by mixing ceramic powders withrganic dispersants to obtained better fluidity of the mixture.rganic–inorganic composites have attracted substantial atten-

ion because of the potential of combining advanced properties

f organic and inorganic components. Formation of compos-te films based on polyelectrolyte and inorganic nanoparticlesas become the subject of extensive experimental work.4 Many

∗ Corresponding author. Tel.: +48 12 639 5104; fax: +48 12 425 1923.E-mail address: ncadamcz@cyf-kr.edu.pl (Z. Adamczyk).

aadspth

955-2219/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.oi:10.1016/j.jeurceramsoc.2006.07.013

mportant studies focused on layer by layer deposition, whichtilized electrostatic interactions between oppositely chargedpecies. There is a growing interest in the use of composite mate-ials in humidity sensors, batteries, microelectronic devices,agnetic memories, and quantum dot devices. Novel compos-

te materials containing inorganic nanopatrticles embedded in aolyelectrolyte matrix are currently under investigation for cat-lytic, optical and other applications.

Ceramic particles can be stabilized in aqueous solutions bylectrostatic and steric interaction. However, it is difficult tobtained homogenous, well-dispersed suspension of high solidontent. Surface chemical properties are influenced by manyactors, such as the nature of the ceramic particles, size distribu-ion, surface impurities, nature of solvent, pH value.3,5–7 In suchhigh solid loading system, the particles have a tendency towardgglomeration. This may increass the viscosity of mixtures andecrease their moldability. The agglomerates can in principle be

tudied by using the electron microscopy techniques (for exam-le scanning electron microscopy SEM). However, it is difficulto identify individual agglomerates in the mixture because ofigh packing density of the powder. Moreover, the examination

2 European Ceramic Society 27 (2007) 2209–2215

ipa

ebhiar

tavotcrawAi

2

2

pA4fpua

b(adczwN

tZtAna

fhmpe

Table 1Physicochemical data of zirconia

Property Value Remarks

Average particle size,standard deviation (�m)

0.64, 0.4 Malvern PCS, pH 10,I = 10−3 M

Specific surface area (m2/g) 5 BETZeta potential (mV) −58 Electrophoresis, pH 10,

I = 10−3 MpH suspension 10pH of the isoelectric point 4.7 Malvern Zetasizer, titration

methodSA

v

mdp

uirvg1e

odtTntwav

dtrztsaU

Zo

2

a

210 B. Jachimska, Z. Adamczyk / Journal of the

s carried out in a gaseous medium or in vacuum, so the sam-le needs to be dried that severely disturbs the real structure ofggregates in suspensions.

Experimental evidence suggested that the degree of agglom-ration of particles in suspensions can be most directly assessedy the dynamic viscosity measurement studied usually for theigh volume concentration range of suspensions.8,9 This lim-ts the accuracy of such measurements because of the lack of anppropriate theoretical model of viscosity for high concentrationange.

Therefore, the goal of this paper is to prove experimen-ally that the structure of ceramic suspensions, in particular thegglomeration degree can be accurately determined from lowolume fraction viscosity measurements. We have chosen forur studies zirconia, which is extensively used in ceramic indus-ry, for producing biocompatible coatings and in the investmentasting industry for producing forms of exceptional temperatureesistance. As the model organic dispersant we have selected thenionic polyelectrolyte:poly(sodium 4-styrenesulfonate) (PSS),hich is stable over a wide range of pH and ionic strength.10

dditionally, its molecular structure, charge and conformationn solutions are well characterized.11

. Experimental

.1. Materials and methods

The zirconia oxide used in this work was a commercialowder (nominal size fraction < 1 �m, purity > 99.5%) fromlfa Aeaser a Johnson Matthey company. Anionic poly(sodium-styrenesulfonate (PSS), Mw = 70 000 g/mol) was purchasedrom Aldrich. Solutions of PSS were prepared by dissolvingolyelectrolytes in deionized water and subsequent filtrationsing the 0.22 �m filter. The pH was adjusted with standardnalytical grade HCl or NaOH solutions.

The specific surface area of the zirconia powder, determinedy means of N2 adsorption using B.E.T. multipoint methodQuantasorb Jr., Quantachrome, USA), was 5.0 m2/g. The phasend crystal structure of zirconia oxide was identified by X-rayiffraction (XRD Philips). The analysis shows that the the zir-onia powder has a monoclinic structure. The estimation ofirconia particle (or aggregate) diameter in aqueous solutionas performed by dynamic light scattering (Malvern Zetasizerano AS).The zeta potential of zirconia suspension was measured by

he microelectrophoretic method using Malvern Zetasizer NanoS apparatus. The ionic strength was adjusted using NaCl solu-

ion. Each zeta potential was averaged from 10 measurements.ll measurement were performed at 293 K. For sake of conve-ience all the physicochemical data characterizing the zirconiare collected in Table 1.

The dynamic viscosity of colloidal zirconia and PSS as aunction of the volume concentration was measured using a

ome made capillary viscometer in vertical orientation. Auto-atic detection of suspension level via electric conductivity

roved very useful for non-transparent suspensions. In eachxperiments, the flow rate time t of the suspension volume

ssaw

pecific density (g/cm3) 5.47, 5.79 Suspension dilution, producerpparent density (g/cm3) 3.04 Viscosity measurements

sus = 10 cm3, through capillary with internal diameter, R, waseasure. The average flow rate was determined from the depen-

ence 〈V 〉 = vsus/πR2t, where t is the time of passing the sus-ension through the capillary.

The temperature of measurements was maintained at 293 K,sing a thermostat. Various capillaries of internal diameter rang-ng from 0.5 to 2 mm were used, depending on the viscosityange. The device was calibrated using pure liquids of knowniscosity, such as, water, butyl and amyl alcohols or ethylenelycol. The accuracy of viscosity determination for the range–10 mPa s was estimated to be 0.5% with the main source ofrror stemming from the suspension density measurements.

The zirconia suspensions of the volume fractions in the rangef 0.005–0.07, were prepared by dispersing zirconia powder inistilled water or in of the appropriate PSS solutions concentra-ion, at a pH adjusted by the standard solution of 0.1 M NaOH.he amount of PSS required to form monolayer on the zirco-ia particles surface was first dissolved in distilled water, thenhe zirconia powder was added to the solution. The suspensionas ultrasonicated for 30 min in order to break up aggregates

nd to remove traces of gases. Every data point of the dynamiciscosity was an average of five measurements.

The amount of polyelectrolyte adsorbed onto zirconia wasetermined using the solution depletion method. After mixinghe zirconia suspension with the polyelectrolyte solution, at theequired concentration (depending on the volume fraction ofirconia powder) under vigorous stirring, the mixture was cen-rifuged at 10,000 rpm for 30 min. The time assured a completeedimentation of the particles. The supernatant was decantednd the concentration of unadsorbed PSS was determined byV spectroscopy at a wavelength of 225 nm.Zeta potential of PSS was determined by using the Malvern

etasizer Nano ZS. The effect of pH was studied for the rangef 3–10.

.2. Physicochemical characteristics of zirconia

The size distribution of zirconia suspensions preparedccording to the above procedure at pH 10 and I = 5 × 10−3 M, is

hown in Fig. 1a. As can be observed, the size distribution of theuspension is rather broad and asymmetrical with the main peakt 0.5 �m. This suggests that a significant fraction of aggregatesas present in the zirconia suspension. The averaged particle

B. Jachimska, Z. Adamczyk / Journal of the European Ceramic Society 27 (2007) 2209–2215 2211

F ) ZrOo icle si

dAca0

t(cBpuodt

3

itautaqetfbptrheJn

eds

Hitculated if the real density of particles in suspension is known.The density was determined by the dilution method, i.e., fromthe dependence of the suspension density on the weight frac-tion of zirconia w = mp/msus, where mp is the particle mass

ig. 1. Particle size distribution of zirconia determined by the PCS method. (af the size 0.40 �m. (b) ZrO2 with PSS at pH 10, I = 5 × 10−3 M, averaged part

iameter was 0.64 �m with the standard deviation of 0.40 �m.similar but more narrow size distribution was obtained for zir-

onia suspension in PSS solution at pH 10 and I = 5 × 10−3 Ms can be seen in Fig. 1b. The averaged particle diameter was.56 �m with the standard deviation of 0.25 �m.

The size of primary particle of zirconia was determined byhe scanning electron microscopy (SEM). From the SEM imagesshown in Fig. 2) it was estimated that the size of primary parti-les was 0.2 �m. It is interesting to compare this value with theET data, giving specific surface area S = 5.0 m2/g as mentionedreviously. By assuming a spherical shape of the particles andsing the formula d = 6/ρpS (where ρp is the specific densityf the solid zirconia, equal 5.79 g/cm3, as specified by the pro-ucer) one obtains d = 0.21 �m. As can be noticed this is similaro the primary particle size derived from the SEM observations.

. Result and discussion

It is known that the stability and aggregation degree of oxides,n particular zirconia, depends to a critical extent on pH valuehat governs the number of hydroxyl groups increasing the neg-tive charge of particles. The increase in the surface chargesually increases the absolute value of the zeta potential of par-icles (the electric potential in the slip plane) which results inn increased stability of suspensions.12 The zeta potential is auantity well accessible experimentally, e.g., by using the micro-lectropheretic method as done in this work. The dependence ofhe zeta potential of ZrO2 suspensionon on the pH determinedor a fixed ionic strength of 10−3 M is shown in Fig. 3. As cane noticed, for pH 4.7, the zeta potential of the suspension ofure zirconia was close to zero. This pH value is referred to ashe isoelectric point. In this respect our data agree well with thateported by Ewais et al.3 and Crucean et al.13,14 On the otherand, Fengqiu et al.5 reported pH 6 as the isoelectric point. Gen-rally, there is a rather large spread in the reported data, e.g.,ohnson et al.6 reported pH 7.5 as the isoelectric point of zirco-ia. This issue is discussed at length in the book of Kosmulski.14

These high values of zeta potential (in absolute terms) arexpected to stabilize of the zirconia suspension. However, theispersion of aggregates seems rather unlikely. This hypothe-is was tested by using the dynamic viscosity measurements.

F(

2 at pH 10, I = 5 × 10−3 M, averaged particle size 0.64 �m, standard deviationze 0.56 �m, standard deviation of the size 0.25 �m.

owever, a prerequisite of a correct interpretation of viscos-ty measurements is the knowledge of the volume fraction ofhe solid in the suspensions ΦV. This quality can only be cal-

ig. 2. SEM image of zirconia powder deposited on mica surface: (a) ZrO2 andb) ZrO2 with PSS.

2212 B. Jachimska, Z. Adamczyk / Journal of the European Ceramic Society 27 (2007) 2209–2215

F

anmaub

ρ

wo

t

ρ

wirbft

Ff

Fig. 5. The dependence of the intrinsic viscosity [η] = ηs/η of ZrO2 suspensionso −3

Tη

5pttf

ipfat0Φ

eE

ig. 3. Zeta potential of ZrO2 as a function of pH determined for I = 1 × 10−3 M.

nd msus is the suspension mass. The weight fraction of zirco-ia was changed within the range 0–0.3 by taking appropriateass of the zirconia powder and the solvent (electrolyte). By

ssuming that the volume of the electrolyte remained the samepon mixing, the density of the zirconia/electrolyte mixture cane expressed by the formula8:

sus = ρe1

1 + w(ρe/ρp − 1)(1)

here ρe is the solvent (electrolyte density) and ρp is the densityf zirconia oxide particles (dry).

From Eq. (1) one can deduce that zirconia particle density inhe suspension can be calculated from the dependence:

p = ρe

1 + tg α(2)

here tg α is the slope of theρe/ρsus versuswdependence (whichs negative in our case) that is expected to be a straight line. The

esults shown in Fig. 4 suggest that these dependencies can welle fitted by a linear regression line with the slope which givesor the zirconia oxide density 5.47 g/cm3. This is slightly smallerhan the value supplied by the producer for solid zirconia, equal

ig. 4. The dependence of ρe/ρsus on the weight fraction w of ZrO2, determinedor pH 10.

η

w

w

ltdpeuia

C

wi

e

n the volume fraction ΦV = wρsus/ρp determined for pH 10, I = 5 × 10 M.he broken line denotes the theoretical results calculated from Einstein formula

s/η = 1 + 2.5ΦV and the solid line represent the linear regression fit.

.79 g/cm3. The deviation is probably caused by the intrinsicorosity of zirconia particles. Some of the pores of primary par-icles forming aggregates were probably closed, inaccessible forhe solvent. By knowing the real density of particles, the volumeraction can be calculated as ΦV = wρsus/ρp.

Typical experimental data presenting the dependence of thentrinsic viscosity [η] = ηs/η (where ηs is the viscosity of the sus-ension and η is the viscosity of the electrolyte) on the volumeraction of zirconia for the low concentration range (ΦV < 0.08)re plotted in Fig. 5. It is interesting to note, that the weight frac-ion of zirconia in these suspensions was much higher attaining.4 (for ΦV = 0.08). As can be seen, the dependence of [η] onV of zirconia can well be fitted by a linear regression. How-

ver, the slope of this line d[η]/dΦV was found larger than theinstein formula predicts, written in the usual form:

¯ = ηs

η= 1 + CVΦV (3)

ere CV is the dimensionless constant.In our case the slope was found equal 4.5, in stead of 2.5,

hich is 1.8 times larger than predicted by the Einstein formula.The deviation from the Einstein formula occurring for the

ow volume fraction of suspensions is interpreted in terms ofhe primary electroviscous effect stemming from shear inducedeformations of ionic double layers surrounding suspendedarticles.15 This leads to increased dissipation of mechanicalnergy resulting in increased viscosity of suspensions over val-es predicted by the Einstein model. This deviation is expressedn terms of the primary electroviscous function, expresseds16–19:

V = 2.5(1 + p) (4)

here p is the primary electroviscous function depending on theonic strength and composition of the electrolyte.

In the case where the diffusion coefficients of ions forming thelectrolyte do not differ too much, one can derive an analytical

Europ

e

p

wζ

tstta

ctoiEtfctm

bavfldhpirv

P

wtd

c

〈

1agldocbarbr

scow

fit(iebsga

pb

ρ

a

ε

iosrobmt

cppcpi7dsurements as a function of the pH and ionic strength. The datacompiled in Table 2 show that the zeta potential of PSS increasedfrom −90 ± 5.0 mV for I = 1 × 10−3 M to −55 ± 5.0 mV forI = 0.15, that suggests a significant reduction in the number of

Table 2Zeta potential ζp of PSS (mV) at pH 10

I (M) Zeta potential (MV)

B. Jachimska, Z. Adamczyk / Journal of the

xpression for p, valid for spherical particles18:

= 2εζ2(1 + κa)2F (κa)

3πηDi

(5)

here ε is the dielectric constant of water (relative permittivity),the zeta potential of the particle forming the suspension, Di

he ith ion diffusion coefficient, F(κa) the function of the dimen-ionless κa parameter, κ−1 = (ε kT/8πe2I)1/2 the double-layerhickness, k the Boltzmann constant, T the absolute tempera-ure, e the elementary charge, I the ionic strength of solutionnd a = d/2 is the particle radius.

For a ionic strength of I = 3 × 10−3 M, κa equals 56 for the zir-onia suspensioun (assuming 0.6 �m as the averaged diameter ofhe aggregate). As calculated in18 for this value of κa, the valuef the function (1 + κa)2F(κa) is about 3 × 10−4. Then, by tak-ng ζ = −50 mV and D = 2 × 10−5 cm2/s one can estimate fromq. (5) that p = 7 × 10−4. As can be seen, this value is more than

hree orders of magnitude smaller than experimentally foundor zirconia, equal 0.8. The comparison proves quite unequivo-ally, that the primary electroviscous effect is not responsible forhe deviation of zirconia suspension viscosity from the Einstein

odel.Therefore, the most probable reason of this deviation can

e sought in the presence of aggregates in the suspension. Forggregates of non-spherical shape, the increase in the intrinsiciscosity of suspensions occurs because of the increase in theuid velocity gradients due to particle rotation. This effect wasescribed quantitatively by Brenner19 for spheroidal particles,aving both a prolate and an oblate shape. The parameter ofrimary significance governing suspension viscosity in this cases the Peclet number (for rotary Brownian motion) defined as theatio of the characteristic rotation velocity 〈G〉d to the diffusionelocity D/d:

e = 〈G〉 d2

4D(6)

here 〈G〉 is the averaged shear rate in the capillary, D = kT/3πηdhe diffusion coefficient of the aggregate and d is the typicaliameter of the aggregate.

In the case of the capillary viscometer the averaged shear ratean be calculated as

G〉 = 8vsus

3πR3t(7)

By taking the typical experimental data: R = 0.035 cm, vsus =0 cm3, t = 50 s, one obtains 〈G〉 = 4 × 103 s−1. Using this valuend taking d = 0.66 �m as the averaged size of the zirconia aggre-ate one can calculate from Eq. (7) that Pe = 6.6 × 102. Such aarge value of Pe indicates that the suspension viscosity wasominated by shear rather than diffusion. Thus, in the limitf such large Peclet numbers, the intrinsic viscosity of parti-le suspensions attains the limiting, high-shear values, tabulatedy Brenner.19 For prolate spheroids having the length to width

spect ratio of 3 the limiting value of [η] = 2.684, for aspectatio 4, [η] = 2.801 and for aspect ratio 5, [η] = 2.918. As cane noticed, even for very elongated spheroids, i.e., for aspectatio 5, the correction to the Einstein’s formula is about 12%. It

1510

ean Ceramic Society 27 (2007) 2209–2215 2213

eems, therefore, that the non-spherical shape of zirconia parti-le aggregates is not responsible for the deviation of the slopef [η] versus ΦV dependence from the from Einstein’s formula,hich was, as previously estimated, about 180%.The most likely explanation is that the zirconia aggregates

orm a loose structure, characterized by a high degree of poros-ty, as suggested by the SEM picture shown in Fig. 2. As a result,he effective hydrodynamic volume of particles in the suspensionat fixed mass fraction) becomes larger than the net volume ofsolated primary particles, which results in a decreased appar-nt density of particles. A similar model was effectively usedefore to explain the anomalous viscosity behavior of silicauspensions.8 It was postulated that upon contact with water, ael-like layer was formed on the surface of silica, having lowerpparent density than the core volume of particles.

Since the primary electroviscous effect remains negligible asreviously estimated, the apparent density of the aggregates cane calculated from the dependence8:

∗ = 2.5ρp

CV(8)

By knowing the apparent density, one can calculated the aver-ge porosity of aggregates εp from the formula8:

p = ρp − ρ∗

ρp − ρe(9)

By substituting into Eq. (9) the data obtained in our exper-ments, i.e., ρp = 5.47 g/cm3, ρ* = 3.03 g/cm3, ρe = 1.01 g/cm3

ne obtains εp = 0.55. Hence, the apparent volume fraction ofolid in the aggregate equals 1 − εp = 0.45. This seems a quiteeasonable estimate in view of the fact that the maximum packingf spheres in three dimensions is 0.62.20 Our value is smallerecause the aggregate structure is more loose than under theaximum packing when, be definition, all particles must con-

act with each other.Having established the basic relationship for the pure zir-

onia suspension, we have studied the effect of the anionicolyelectrolyte PSS on the viscosity of zirconia suspension. Thisolyelectrolyte, bearing sulfonic groups in chain was thoroughlyharacterized in our previous work.11 The extended length of theolyelectrolyte chain, having the molecular weight of 70,000s 91 nm and the length to diameter ratio of the chain equals8. One can draw quantitative conclusions about the ionizationegree of the PSS chain from the electrophoretic mobility mea-

× 10−3 −90 ± 5.0× 10−3 −83 ± 5.0× 10−2 −74 ± 5.0.15 −55 ± 5.0

2214 B. Jachimska, Z. Adamczyk / Journal of the Europ

FZ

ittitam

fsccdIpisctvc

Ff2

mosltPzct

[

w

p(fAe

tiipedcaactmts

ig. 6. Zeta potential as a function of pH determined for I = 1 × 10−3 M. (�)rO2 with PSS (curve 1); (�) PSS (curve 2).

onic groups at the PSS molecule surface. On the other hand,he dependence of the zeta potential of PSS on pH of the solu-ion measured for a fixed ionic strength I = 1 × 10−3 M is shownn Fig. 6. As can be seen, for the pH range studied, i.e., 5–10he zeta potential of PSS remains fairly constant that suggestspH- independent ionization degree of surface groups of thisolecule.The result of viscosity measurements of PSS solutions per-

ormed for PSS volume fraction range ΦV < 0.004 and the ionictrength of I = 5 × 10−3 M and 0.15 M are presented in Fig. 7. Asan be seen, the dependence of [η] on ΦV for both ionic strengthan well be fitted by a linear regression. The slope of these lines[η]/dΦV was found equal to 50 for I = 5 × 10−3 M and 25 for= 0.15 M. This is again much higher than the Einstein formularedicts for spherically shaped particles. It seems that a propernterpretation of these data can be sought in the highly non-pherical shape of PSS molecules in solutions. Again, these data

19

an be interpreted in terms of the Brenner model. However, inhe case of PSS, the rotary Peclet number assumes much smalleralues than for zirconia because of much higher diffusion coeffi-ient of PSS. By taking 〈G〉 = 4 × 103 s−1, L = 91 nm (extended

ig. 7. The dependence of the intrinsic viscosity of PSS [η] = ηs/η on its volumeraction ΦV, pH 6.5–10, (�) I = 5 × 10−3 M (curve 1) and (�) I = 0.15 M (curve).

i

zTwDInonCfsnaazash

r

ean Ceramic Society 27 (2007) 2209–2215

olecule length) and D = 1.8 × 10−7 cm2/s (for I = 5 × 10−3 M)ne can calculate from Eq. (7) that Pe = 0.46. For higher ionictrength this should be definitely smaller because of the smallerength of the molecule. Since this value is smaller than unity,he suspension viscosity is dominated by the rotary diffusion ofSS molecules rather than by the shear as was the case for theirconia suspension. In this limit, of diffusion dominated vis-osity Brenner19 derived the approximate formula describinghe intrinsic viscosity of prolate spheroid polymer solution:

η] = L2

15

[3

ln 2L − 0.5+ 1

ln 2L − 1.5

]+ 8

5(10)

here L = L/d is the ratio of the length to width of the spheroid.It can be calculated from Eq. (10) that for L = 78 (the value

ertinent to the extended PSS molecule) [η] = 383, for L = 50the molecule bend to the form of a semi-circle) [η] = 177 andor L = 25 (the molecule bend to the form of a circle) [η] = 55.s can be seen, the latter intrinsic viscosity agrees well with the

xperimental value obtained for the ionic strength of 5 × 10−3.After measuring the viscosity of zirconia and PSS separately,

he viscosity of mixtures was studied in order to estimate themprovement in the fluidity upon the addition of PSS. The flu-dity can be increased by increasing the dispersion degree of theowder promoted by adsorption of PSS. The amount of poly-lectrolyte adsorbed on zirconia was determined by the solutionepletion method. It was found that irrespective on the initialoncentration of PSS in the suspension (varied between 1000nd 5000 ppm) the amount of PSS adsorbed on particles waslmost constant and equal 2.5 × 10−3 g per one gram of the zir-onia powder (this corresponds to 5 × 10−4 g per square meter ofhe dry powder). By taking the affective cross-section of the PSS

olecule as 91 × 11 nm (1.06 × 10−16 m2) one can estimate thathe amount of adsorbed PSS corresponds to the dimensionlessurface coverage of 0.45, which seems a quite reasonable valuen view of previous estimates.21

Adsorption of PSS on zirconia decreased significantly theeta potential of particles for the entire range of pH (see Fig. 6).his is fully in accordance with the results of Fengqiu et all.5

ho determined zeta potential of zirconia powder in solutions ofarvan C (an anionic polyelectrolyte with carboxylic groups).

t was observed that for pH > 3 zeta potential of zirconia becameegative upon adsorption of Darvin C attaining a minimum valuef −60 mV for pH 8. It is not obvious what is the physical mecha-ism of adsorption of anionic polyelectrolytes (PSS and Darvin) on negatively charged zirconia surface, which is the case

or pH > 5 as can be deduced from the data shown in Fig. 3. Iteems that a possible explanation of this fact is the heteroge-eous nature of zirconia particles dispersed in solution. Hence,dsorption of PSS may occur on surface areas bearing locallypositive charge in the nanoscale, although the net charge of

irconia surface, in the microscale, remains negative. Certainly,unequivocal confirmation of this hypothesis requires further

ystematic studies, which are outside the scope of this work,owever.

It is interesting to mention that, in contrast to the resultseported by Fengqiu et al.5 in our case, adsorption of PSS on zir-

B. Jachimska, Z. Adamczyk / Journal of the Europ

Fig. 8. The dependence of the intrinsic viscosity [η] = ηs/η ZrO2 vs. the volumefraction ΦV = wρsus/ρp, pH 10, I = 5 × 10−3 M. (�) ZrO2, the line denotes thels(

casPtPtcd3fpooatfe

4

cwTmeachbmf

psvTAPc

Ptosvc

A

k

R

1

1

1111

1

1

112

inear regression [η] = 1 + 4.5 ΦVZrO2(curve 2); (©) ZrO2 + PSS mixture, the

olid line denoted the product of intrinsic viscosities [η] = [η]ZrO2+ [η]PSS =

1 + 4.5 ΦVZrO2) + (1 + 50 ΦVPSS ) (curve 1).

onia did not increase too significantly the dispersion degree ofggregates. This is evident from the particle size measurementshowing that the average size of zirconia particles dispersed inSS (at pH 7.4) was 0.56 �m, which does not differ much from

he value observed in PSS free solutions. The limited influence ofSS on the dispersion degree of zirconia also was confirmed by

he viscosity measurements whose results are shown in Fig. 8. Asan be seen, the dependence of the intrinsic viscosity of zirconiaispersed in PSS solution (variable polymer concentration from50 ppm to 2000 ppm, pH 10, I = 5 × 10−3 M) on the volumeraction solid can well be fitted by a line with the slope being theroduct of the viscosity of zirconia dispersed in PSS free solutionf the same composition and the viscosity of PSS alone. Hence,ur results seem to confirm the effective medium hypothesis22

ccording to which the relative viscosity of a suspension havinghe solid concentration ΦV1 + ΦV2 (where ΦV1 is the volumeraction of zirconia and ΦV2 is the volume fraction of PSS)quals to the product of relative viscosities η(ΦV1 )η(ΦV2 ).

. Concluding remarks

It has been demonstrated experimentally that the relative vis-osity of dilute zirconia suspensions increased more rapidlyith the volume fraction that the Einstein formula predicted.his discrepancy could not be explained in terms of the pri-ary electroviscous effect because of small thickness of the

lectric double-layer. As a plausible explanation the presence ofggregates in the zirconia suspension was given, which wereomposed of primary particles of the size ca. 200 nm. The

ypothesis of the presence of aggregates was also confirmedy the SEM. The viscosity measurements allowed one to deter-ine the averaged solid content in these aggregates, which was

ound close to 0.45.

22

ean Ceramic Society 27 (2007) 2209–2215 2215

Analogous measurements of the viscosity of the anionicolyelectrolyte-PSS used to enhance the stability of zirconiauspensions showed that the slope of the intrinsic viscosity vs.olume fraction attained values as high as 50 for I = 5 × 10−3 M.his effect was interpreted in terms of the flexible rod model.ccording to this model, for higher electrolyte concentration, theSS molecule assumes the shape of a semi circle or sinusoidalurve rather than a random coil structure as often suggested.

It was further confirmed experimentally that adsorption ofSS on zirconia decreased significantly the negative zeta poten-

ial but did not influence too significantly the dispersion degreef zirconia particles. Also the viscosity of the zirconia suspen-ion in PSS solutions was found equal to the product of theiscosity of zirconia dispersed in PSS free solution of the sameomposition and the viscosity of PSS alone.

cknowledgements

This work was supported by MNiSW Grant 4T08B 03425.The help of Ms. K. Kusak in performing the experiments is

indly acknowledged.

eferences

1. Liu, D. M., Ceram. Int., 2000, 26, 279.2. Baklouti, S. M., Romdhane, R. B., Boufi, Pagnoux, S. C., Chartier, T. and

Baumard, J. F., J. Eur. Ceram. Sci., 2003, 23, 905.3. Ewais, E., Zaman, A. A. and Sigmund, W., J. Eur. Ceram. Sci., 2002, 22,

2805.4. Decher, G., Multilayer Thin Films. Wiley-VCH, 2002.5. Fengqiu, T., Xiaoxian, H., Yufeng, Z. and Jingkun, G., Ceram. Int., 2000,

26, 93.6. Johnson, S. B., Franks, G. V., Scales, P. J., Boder, D. V. and Healy, T. W.,

Int. J. Miner. Process., 2000, 58, 267.7. Piconi, C. and Maccauro, G., Biomaterials, 1999, 20, 1.8. Adamczyk, Z., Jachimska, B. and Kolasinska, M., J. Colloid Interface Sci.,

2004, 273, 668.9. Okada, K. and Nagase, Y., J. Chem. Eng. Jpn., 2000, 33, 168.0. Atkins, A., Physical Chemistry. Oxford University Press, Oxford, 1994, p.

C18.1. Adamczyk, Z., Zembala, M., Warszynski, P. and Jachimska, B., Langmuir,

2004, 20, 10517.2. Hunter, R., Foundations of Colloid Science. Oxford, 2001.3. Crucean, E., Rand, B. and Brit, T. J., Ceram. Soc., 1979, 78, 96.4. Kosmulski, M., Chem. Properties Mater. Surf.: Surf. Sci. Ser., 2001, 102.5. Masiliyah, J. H., Neale, G., Malysa, K. and van de Ven, T. G. M., Chem.

Eng. Sci., 1987, 42, 245.6. Honig, E. P., Punt, W. F. J. and Offermans, P. H. G., J. Colloid Interface Sci.,

1990, 134, 169.7. Adamczyk, Z., Siwek, B. and Zembala, M., Bull. Pol. Acad. Chem., 2000,

48, 231.8. Laven, J. and Stein, H. N., J. Colloid Interface Sci., 2001, 238, 8.9. Brenner, H., Int. J. Multiphase Flow, 1974, 1, 195.0. Greenwood, R., Luckham, P. F. and Gregory, T., J. Colloid Interface Sci.,

1997, 191, 11.1. Stoll, S. and Chodanowski, P., Macromolecules, 2002, 35, 9556.2. Pal, R., Rheology of emulsions containing polymeric liquids. In Encyclope-

dia of Emulsion Technology, Vol. 4, ed. P. Becher. Marcel Dekker, Newyork,Basel, Hong Kong, 1996.